Code obj support

[clubby789]

For both tools:

• Adds a boolean -c flag and a string -v paramater for specifying 'this file is a code object' and the Python version respectively
• Adds a new method to load a PycModule from a code object (skips header loading and manually sets magic from major/minor)
• Add an INVALID variant to the PycMagic enum for an invalid magic value

[zrax]

setVersion() already sets these... Additionally, since we're parsing the version directly, we may want to have a way to indicate that Unicode mode is being used for Python 1.6-2.7. Perhaps checking for a 'u' suffix to the version number would be sufficient?

[clubby789]

Isn't the CO_FUTURE_UNICODE_LITERALS flag set within the code object itself, and wouldn't need to be explicitly added?

[zrax]

The unicode bytecode magic number is used when Python 2.x is run with the (undocumented) -U flag, and is not related to the CO_FUTURE_UNICODE_LITERALS code flag -- the two can be present or absent independently of each other...

However, from looking at the code, it looks like the only place we actually use m_unicode is when printing the Python version in the comments... There may be something missing there that will need to be addressed eventually, but since this flag was rarely used and doesn't have any meaningful impact on the output, I suppose this can be ignored for now.